CVPR SignEval2026 Models
Model checkpoints for the S23 super-ensemble that won the CVPR 2026 Isolated Sign Language Recognition (ISLR) challenge.
Model source
- Kind
- Other
- File size
- 21.5–107 MB
- Version
- vRevision 9bef669
- Updated
- Sep 17, 2026
- Downloads
- 0
- Source
- Hugging Face
Source excerpt
Model checkpoints for the S23 super-ensemble that won the CVPR 2026 Isolated Sign Language Recognition (ISLR) challenge.
Sources
1 source
Sources
1 sourceVerified Sep 2
Download files
Model artifacts
64 artifacts
Model artifacts
64 artifactsr15_b2_cm_trip_unified/fold0/best.pt
pt · 30.5 MB · SHA-256 94f5ce6452da…4aef · Hugging Face
Download filer15_b2_cm_trip_unified/fold1/best.pt
pt · 30.5 MB · SHA-256 de4bd17fb910…4595 · Hugging Face
Download filer15_b2_cm_trip_unified/fold2/best.pt
pt · 30.5 MB · SHA-256 adb1aba47db3…33ff · Hugging Face
Download filer15_b2_cm_trip_unified/fold3/best.pt
pt · 30.5 MB · SHA-256 11ff5e7eac8f…ce6b · Hugging Face
Download filer15_b2_cm_trip_unified/fold4/best.pt
pt · 30.5 MB · SHA-256 8be3b550fabe…b8b3 · Hugging Face
Download fileSource excerpts
2 excerpts
Source excerpts
2 excerpts--- license: mit tags: - image-classification - isolated-sign-language-recognition - sign-language - ensemble - cvpr library_name: pytorch --- # CVPR-SignEval2026 — S23 winning ensemble (🥇 1st place) Model checkpoints for the **S23 super-ensemble** that won the CVPR 2026 Isolated Sign Language Recognition (ISLR) challenge. > **🏆 S23: `0.93101` private / `0.92165` public — 1st place.** > 17 configs × 5 folds across **4 CNN families / 5 backbone variants** (EfficientNet-B1/B2, > ConvNeXt-tiny, MobileOne-S2, RegNetY-008), combined by CV-weighted probability averaging. - **Code / training / inference:** https://github.com/CristianLazoQuispe/CVPR-SignEval2026 - **License:** MIT ## What's in this repo 83 fold-level checkpoints (`best.pt`), organized as `<config>/fold<k>/best.pt`, plus each config's `config.yaml` and `experiment_summary.json`. (17 configs × 5 folds = up to 85; the official S23 used **83** — `r15_regy_cm` contributed 3/5 folds.) | Backbone | Configs | |:--|:--| | EfficientNet-B1 | `r9_both_trip` · `r9_cutmix_ls02` · `r9_cutmix_p08` · `r9_cutmix_a20` · `r12_01`…`r12_05` · `r15_b1_s43` | | EfficientNet-B2 | `r15_b2_cm_trip_unified` | | ConvNeXt-tiny | `r9_cnxt_trip` · `r15_cnxt_cm` | | MobileOne-S2 | `r13_mobileone` · `r14_mob_cm2_trip` | | RegNetY-008 | `r13_regnety` · `r15_regy_cm` | ## Results — ensemble scaling (Kaggle private LB) | Submission | Configs | Models | Architectures | Private LB | |:--|:--:|:--:|:--:|:--:| | S15 | 6 | 30 | 1 | 0.92612 | | S16 | 4 | 20 | 1 | 0.92816 | | S17 | 9 | 45 | 1 | 0.92897 | | S22 | 15 | 75 | 4 | 0.93019 | | **🥇 S23** | **17** | **83** | **5** | **0.93101** | ## Usage Download all checkpoints and run inference + CV-weighted ensembling with the code repo: ```python from huggingface_hub import snapshot_download local = snapshot_download(repo_id="CristianLazoQuispe/CVPR-SignEval2026-models", repo_type="model") # -> place under experiments/ and run scripts/build_s23.sh (see code repo) ``` Each checkpoint is a standard PyTorch `state_dict` saved with `torch.save`. The matching architecture/preprocessing for every config lives in its `config.yaml`. See the code repo's README ("Reproduce the winning ensemble") for the end-to-end inference → ensemble pipeline. ## Reproducibility All models trained with `seed: 42`, `deterministic: true`, `amp: false` for bit-reproducible runs. ## Citation See...
Source context: 0 downloads · 0 likes · Pipeline image-classification · Library pytorch · Repo CristianLazoQuispe/CVPR-SignEval2026-models
